piabrota← proof index

Contextual bandit: model and decision layers

Contextual bandit: model and decision layers An architecture diagram generated by Archify. Member · opens the homepage · Architecture component Member opens the homepage Context c · history, country · Architecture component Context c history, country RewardModel · belief b · Architecture component · probabilistic RewardModel belief b probabilistic DecisionRule · greedy, ε, UCB, TS · Architecture component DecisionRule greedy, ε, UCB, TS Homepage · shows arm a · Architecture component Homepage shows arm a Z3 spec · proved invariants · Architecture component Z3 spec proved invariants Event log · (c, a, r, P(a)) · Architecture component Event log (c, a, r, P(a)) request estimate mean, width arm a proves log reward r → update Legend Request and page Decision rule Model Proof Log Actor